Origin Energy, an Australian-based integrated energy company, is working with Google Cloud and Accenture to create a tool that utilizes satellite imagery, 3D data, visual AI and advanced analytics to remotely help customers understand their solar power needs and opportunity. 

Origin Energy has already helped more than 110,000 Australian homes go solar since it launched its rooftop solar offering over 15 years ago - but given the energy crisis globally, particularly the alarming impact of climate change, Origin is looking at ways to speed up the process of providing information to customers as they weigh up their decision. 

Speaking at Google Cloud’s annual user even this week, Duncan Permezel, General Manager of Retail Sales & Marketing at Origin Energy, said: 

At Origin our purpose is getting energy right for our customers, community and planet. And our ambition and strategy is to lead the transition to net zero through cleaner energy and customer solutions. One way we will do this, and a key pillar of our strategy, is by offering unrivalled customer solutions. 

And part of this is making it simple and easy for customers to access clean and smart energy solutions. 

Historically, Origin made sales of rooftop solar solutions by physically visiting a site, reviewing the customer’s site, and then going away to consider the right system to establish a quote. The customer would then receive the quote for review and weigh up their options. If the sale progressed, an installation would be scheduled on another day. Permezel said: 

It was often the same person or business doing the sale and then coming back to do the install later. But at Origin, as a national retailer, we do most of our work over the phone - probably around 80%. 

Permezel explained that whether a customer chooses to invest in solar, besides from the environmental benefit, is often down to whether it not they will save money or not. He said: 

Due to the nature of the upfront payment and the benefits over time, the decision for consumers is more like a total cost of ownership decision that has payback periods of typically between three to seven years, if the system is sized correctly.

A consumer’s decision is based on three key drivers. Firstly, the upfront cost of the system, noting any financial or payment terms. The second is the displacement of green energy, by using the energy from the solar system directly in the home. And the third is the earned revenue from energy exported to the grid by, what we call feed in tariffs. 

Permezel said that a household that consumed, for example, five megawatt hours per year from the grid, presolar, may move to a household that only consumes three and a half megawatt hours directly from the grid and produces two megawatt hours from their solar instal - and so offsets one and a half megawatt hours of usage in the home and exports the other half a megawatt hour to the grid. 

They could then be impacted by a couple of other variables. For example, one variable is grid tariffs, where many consumers pay Time Of Use tariffs, which means that the grid price is normally cheapest during the day, when the sun is shining, and the load on the network and the generation is at its lowest. 

The price then becomes higher in the late afternoon to evening when there is more demand on the network from households. 

Another variable is usage profiles - or how much solar generation is occurring from the solar system during sunshine hours. And then there's the differential in rates. Depending on the time of use tariffs, grid rates can sometimes be up around 25 to 30 cents during peak times. They can also be down as low as six to 10 cents per kilowatt hour in off peak times. 

Permezel said that this is where the tool that has been built with Google Cloud and Accenture is proving particularly valuable - helping customers better understand their opportunity from shifting to solar. He explained: 

All of this means there's much more value for the household in the energy that is used directly in the home. So the new system allows users to provide a simple address to look up and see what their solar opportunity is. It has an intuitive interface that allows customers to see in near real time, what the solar setup can look like on the property. 

The inputs utilized include the home, its orientation, the roof size, the roof pitch, and roof material. It also matches up the technical options around the panel types, sizes, the inverter capacity, and even the battery and storage options if that's a viable investment for the customer. 

Given our internal information around the customer's usage, we can also orchestrate an outbound campaign that matches the customer's existing tariffs and usage profiles to the insight about the house. This will allow us to show a customer what the best option for them is from a payback perspective. 

A new way ahead

Origin’s old way of implementing solar required investing a lot of time into each system, with a large amount of human intervention, which also involved gathering information and data from numerous sources. Permezel said: 

Sometimes this was onerous on the customer too. With the new system that we developed, together with Accenture and Google Cloud, we can digitally capture the required information. We then use artificial intelligence and machine learning to optimize the system and options. This then allows us an immediate payback to the customer with a quote.

The machine learning and AI models used means the customer has a tailored solution and they can be confident that the system hasn't been oversized or undersized for their usage profile. 

Historically, assessing the suitability of a roof for solar and determining where to place panels has been performed by knowledge workers, who pieced together many components to arrive at a recommendation. Whilst this is a proven approach, it does take time and it can't really scale to provide real time advice to many customers simultaneously. 

Using this data driven, automated approach, Origin is able to scale more efficiently. Permezel added: 

The innovative Origin model uses visual AI and geometry to enable almost unlimited parallel solution assessments, within a very easy to understand customer experience. There are five key steps in the new process. The first step looks to understand the type of roof being considered. Is it the right material, or is it tiles which could be problematic for the installation? 

The second and third steps break down the overall roof area into segments, provide insights on the roof slope and orientation to the sun, and provide the gross available area that potentially could be used for solar. 

The final steps place panels on the roof in line with the recommendations for the customer in terms of usage, and what is possible from a solution perspective. Collectively, these steps are performed in under 30 seconds. 

Origin is now seeing customers’ journey time shortened while still being able to configure a system and outcome that is personalized for the customers property and usage profile. Permezel said that the quote is accurate and allows confidence in teh assessment of the data and information to build the recommendation quote. He added: 

We are getting higher interaction net promoter scores through the new platform than previously and the customers can choose what time of day they choose to interact with us and consider the purchase options. We're also seeing lower cancellation rates on orders due to the increased confidence in the system design and restructure information. 

This lower cancellation rate leads to more efficient operations, the machine learning and the team progressing the platform are constantly working on improving the proposition and building continuous improvement cycles that have us enjoying the opportunity of many ideas and options to keep refining the performance of the platform and the customer experience.